My friend Joshua Israelson, a former
editor and analog-circuit expert reminded me to explain that the drift mentioned for an auto-zero amplifier (AZA) in the July 14 column (http://rbi.ims.ca/5719-529) relates to thermal drift of the offset voltage for which an AZA provides first-order compensation.
But an AZA cannot compensate for measurement errors caused by sensors. In a Wheatstone-bridge circuit that uses a single strain gauge — really a strain sensor — to measure force, temperature changes slightly alter the sensor's resistance and thus the bridge output. Every conductor has a temperature coefficient of resistance that indicates how resistance changes with temperature.
If ambient temperature remains constant during a measurement, you can manually null out an offset due to temperature. Electronic weighing scales, for example, can quickly compensate for thermal errors between measurements. But if the temperature of the sensor changes during measurements, you need another way to account for thermal errors. (Think of strain gauges applied to a truck chassis for outdoor testing during day and night). Because the sensor constantly measures strain, you never have time to take an “at rest” reading that would let you compensate for thermal errors.
You can put a second strain gauge (SG2) in the bridge circuit to help compensate for thermal errors (see figure, above). The resistance of the added sensor changes with temperature to the same extent as the resistance of the original sensor (SG1), which keeps the bridge balanced. You must place strain gauge SG2 close to SG1 to ensure exposure to the same ambient conditions and mount the reference strain gauge so it experiences no strain. It serves solely as a reference and you don't want to complicate things by including another error source — strain in the reference sensor — in your measurements.
Because you mount strain gauge (SG2) near the original sensor (SG1), in some cases you can put the reference gauge to work. Suppose you need to measure the strain in a piece of steel as it bends during a test. Mount strain gauge SG1 on the side of the steel that will stretch slightly and mount the other strain gauge on the side that will compress slightly. Each sensor will experience strain — compression or expansion — in opposite directions, which will increase the bridge output voltage. And, the two strain gauges experience the same thermal condition which offsets thermal errors.
Keep in mind that materials to which you affix strain gauges expand or contract with temperature changes. So you might have to account for small changes in output signals due to the thermal expansion or contraction of the materials on which you mount strain gauges.
“The Strain Gauge,” Omega Engineering: http://rbi.ims.ca/5719-530
“How They Work: The Strain Gauge,” SensorLand: http://rbi.ims.ca/5719-531
“Measuring Strain with Strain Gauges,” National Instruments: http://rbi.ims.ca/5719-532
|Jon Titus, a former designer and chief editor of EDN and Test & Measurement World magazines, remembers when “fast” signals operated at 10 MHz and programs came on paper tape.